The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the nvm installation script from the official nvm-sh GitHub repository and performs live API queries to well-known services (api.osv.dev, api.github.com, endoflife.date, and registry.npmjs.org) for dependency auditing and support-status checks.
[COMMAND_EXECUTION]: The skill invokes a wide array of legitimate development and analysis utilities, including package managers (pnpm, npm, yarn, bun), compiler tools (tsc), linters (eslint), and test suites (vitest, jest, playwright, cypress) to perform its health audit functions.
[INDIRECT_PROMPT_INJECTION]: The skill operates by analyzing potentially untrusted content from the repository it is auditing.
Ingestion points: The skill reads source files, configuration manifests, and documentation throughout the target repository, as specified in references/repository-inventory.md and references/code-quality.md.
Boundary markers: The instructions do not define specific delimiters or escaping mechanisms to separate the audited repository data from the agent's core instruction set.
Capability inventory: The agent has access to Bash for command execution, the Agent tool for managing sub-tasks, and Write for report generation.
Sanitization: The skill relies on behavioral descriptions and 'Evidence-Based' reasoning rather than explicit technical sanitization of the data it processes.

nextjs-health-audit